Multiple sparse volumetric priors for distributed EEG source reconstruction

Heat transfer due to electroconvulsive therapy: Influence of anisotropic thermal and electrical skull conductivity

Objective. In electroencephalography (EEG) measurements, the signal of each recording electrode is contrasted with a reference electrode or a combination of electrodes. The estimation of a neutral reference is a long-standing issue in EEG data analysis, which has motivated the proposal of different re-referencing methods, among which linked-mastoid re-referencing (LMR), average re-referencing (AR) and reference electrode standardization technique (REST). In this study we quantitatively assessed the extent to which the use of a high-density montage and a realistic head model can impact on the optimal estimation of a neutral reference for EEG recordings. Approach. Using simulated recordings generated by projecting specific source activity over the sensors, we assessed to what extent AR, REST and LMR may distort the scalp topography. We examined the impact electrode coverage has on AR and REST, and how accurate the REST reconstruction is for realistic and less realistic (three-layer and single-layer spherical) head models, and with possible uncertainty in the electrode positions. We assessed LMR, AR and REST also in the presence of typical EEG artifacts that are mixed in the recordings. Finally, we applied them to real EEG data collected in a target detection experiment to corroborate our findings on simulated data. Main results. Both AR and REST have relatively low reconstruction errors compared to LMR, and that REST is less sensitive than AR and LMR to artifacts mixed in the EEG data. For both AR and REST, high electrode density yields low re-referencing reconstruction errors. A realistic head model is critical for REST, leading to a more accurate estimate of a neutral reference compared to spherical head models. With a low-density montage, REST shows a more reliable reconstruction than AR either with a realistic or a three-layer spherical head model. Conversely, with a high-density montage AR yields better results unless precise information on electrode positions is available. Significance. Our study is the first to quantitatively assess the performance of EEG re-referencing techniques in relation to the use of a high-density montage and a realistic head model. We hope our study will help researchers in the choice of the most effective re-referencing approach for their EEG studies.

This paper presents finite-difference-time-domain (FDTD) calculations of the specific absorption rate (SAR) in realistic head models from exposure to a generic handset working as 1750 Mhz. The head models with different sizes were obtained from the same whole-body model. The purpose of this work is to study whether there is a variation of SAR absorption in the same brain, but the sizes of head models are different. The obtained peak SARs in each of the tissues were averaged over 10 grams of tissue in the shape of cube. It was found that the SAR absorption in human brain is dependent on the size of the head model. The induced SAR in brain tissues in smaller head model is larger than that in larger head model. It suggested that the mobile phone dosimetric analysis in most of the previous literatures which only considered head model in the simulation may overestimate the brain exposure compared to the practical situation that the whole-body exposed to the fields radiated by the mobile phone.

By the use of three different head models in EEG dipole analysis, we tried to model the origin of interictal and ictal epileptic activity as precisely as possible. Further, as a control, a second evaluation was made by an independent group to control for interindividual reliability of the dipole source analysis. With the realistic head model (CURRY) considering cortex, skull, and skin segmentation, the spike source was located. In five patients with mesial temporal epileptogenesis, confirmed by successful epilepsy surgery, the spike source was close to the hippocampus, with a mean distance of the dipole source from the hippocampus of 13.6 mm (range, 9-17.2 mm). In one case the ictal EEG also could be analyzed and resulted in a dipole-source localization comparable to the interictal source. In both head models using either pure cortex segmentation only or a concentric three-shell model, the dipole source was systematically dislocated in a more superior position. Data analysis by a second group with independently chosen EEG samples and identical individual head model resulted in deviations of <5.3 mm. Data analysis using independently selected spikes and independently segmented head models resulted in deviations < or =16.7 mm. In four cases of extratemporal epileptogenesis, the origin of interictal epileptiform discharges was localized to the suspected primary epileptogenic zone.

The real head model has been used for the accurate scalp potential modeling. The realistic shapes of head tissues were derived from a set of 2-D magnetic resonance images (MRI) by extracting surface boundaries for the major tissues such as the scalp, the skull, the cerebrospinal fluid (CSF), the white matter, and the gray matter. From boundary data a 3-D volume generic head model has been constructed and a mesh for an arbitrary complexity head shape has been generated for finite-element method (FEM) modeling. This paper first addresses the use of this realistic finite elements head model to solve the EEG forward problem. The accuracy and computational time of the potential modeling are then examined.

The human head is a significant part of the human body, with which we can recognize individuals from a vast universe of populations. Since the early 1970s, considerable effort has been devoted to computeraided modeling of human head for applications varied from realistic effect in computer graphics to custom model generation in modern manufacturing industries. However, realistic head modeling is still a challenge and continues to fascinate computer graphics researchers. The greatest difficulties in human head modeling result from the extremely complex geometric form of the human head. To generate realistic individualized models, most proposed head modeling techniques use the common approach of deforming a generic model into an individualized one, based on individual head information. There are various sources to obtain individual information, e.g., anthropometric data, range data and 2D pictures. Given the individual information, the quality of the resulting individualized model depends on the quality of the generic model and the deformation technique used. In our study, we observe that the mathematical form in which the generic model is represented strongly determines the deformation effect and, thus, determines the quality of the resulting individualized models. During the past few years, the multiresolution modeling technique has been demonstrated to be a powerful tool for highly detailed sculptured object modeling. In our work, we adopt the multiresolution technique for generic head modeling and propose a feature-based deformation technique. We show that using our technique we can generate highly realistic individualized head models with speed and efficiency. We also demonstrate that our proposed technique can result in great efficiency for a wide range of downstream applications.

A finite-element reciprocity solution for EEG forward modeling with realistic individual head models

The light propagation in the head models has been investigated to deduce the volume of the brain tissue sampled by NIR instrument. The light propagation in the adult head is strongly affected by the presence of low scattering cerebrospinal fluid (CSF) layer. Although the brain surface is folded with sulci filled with the CSF, the brain surface in the previous head models has had simple geometry such as a flat or curved surface. In this study, the light propagation in a realistic adult and neonatal head models of which geometry is generated from MRI scan of the head is predicted by Monte Carlo method. The adult neonatal head models consist of small square elements to represent complex geometry and each element specified its scattering and absorption coefficients. The delta-scattering algorithm is implemented to keep the computation time within reasonable limit. The spatial sensitivity profiles for various source-detector spacing are predicted and the difference in the effect of the CSF on the spatial sensitivity profiles in the adult and neonatal head models is discussed. The low scattering CSF affects the light propagation in the adult head model while the distortion in spatial sensitivity profiles in the neonatal head models is not significant.© (2000) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Electroencephalography (EEG) forward solution computes the electrical potentials over the electrodes that are lying on the scalp surface. Forward solution requires a numerical head model to work on. The head model can be spherical or a realistic one. Accuracy of the forward problem can be improved by using realistic head models. Triangulated realistic head model is an input for the Boundary Element method (BEM) which solves the EEG forward problem. In this paper, a software is introduced which constructs 3-D triangulated realistic head models from the Magnetic Resonance images.

Caudal cingulate cortex involvement in pain processing: an inter-individual laser evoked potential source localisation study using realistic head models

The performance of the three-shell spherical head model versus the performance of the realistic head model is investigated when solving the inverse problem with a single dipole, in the presence of noise. This is evaluated by inspecting the average dipole location error when applying a spherical and a realistic head model, for 1000 noisy scalp potentials, originating from the same test dipole and having the same noise level. The location errors are obtained utilizing a local linearization, which is validated with a Monte Carlo simulation. For 27 electrodes, an EEG epoch of one time sample and spatially white Gaussian noise we found that the importance of the realistic head model over the spherical head model reduces by increasing the noise level.

Various source localization techniques have indicated the generators of each identifiable component of movement-related cortical potentials, since the discovery of the surface negative potential prior to self-paced movement by Kornhuber and Decke. Readiness potentials and fields preceding self-paced finger movements were recorded simultaneously using multichannel electroencephalography (EEG) and magnetoencephalography (MEG) from five healthy subjects. The cortical areas involved in this paradigm are the supplementary motor area (SMA) (bilateral), pre-SMA (bilateral), and contralateral motor area of the moving finger. This hypothesis is tested in this paper using the dipole source analysis independently for only EEG, only MEG, and both combined. To localize the sources, the forward problem is first solved by using the boundary-element method for realistic head models and by using a locally-fitted-sphere approach for spherical head models consisting of a set of connected volumes, typically representing the scalp, skull, and brain. In the source reconstruction it is to be expected that EEG predominantly localizes radially oriented sources while MEG localizes tangential sources at the desired region of the cortex. The effect of MEG on EEG is also observed when analyzing both combined data. When comparing the two head models, the spherical and the realistic head models showed similar results. The significant points for this study are comparing the source analysis between the two modalities (EEG and MEG) so as to assure that EEG is sensitive to mostly radially orientated sources while MEG is only sensitive to only tangential sources, and comparing the spherical and individual head models.

Neuroelectromagnetic Forward Head Modeling Toolbox

Computational head and brain volume conductor modeling is a practical and non-invasive method to investigate neuroelectrical activity in the brain. Anatomical structures included in a model affect the flow of volume currents and the resulting scalp surface potentials. The influence of different tissues within the head on scalp surface potentials was investigated by constructing five highly detailed, realistic head models from segmented and processed Visible Human Man digital images. The models were: (1) model with 20 different tissues, that is, skin, dense connective tissue (fat), aponeurosis (muscle), outer, middle and inner tables of the scalp, dura matter, arachnoid layer (including cerebrospinal fluid), pia matter, six cortical layers, eye tissue, muscle around the eye, optic nerve, temporal muscle, white matter and internal air, (2) model with three main inhomogeneities, that is, scalp, skull, brain, (3) model with homogeneous scalp and remaining inhomogeneities, (4) model with homogeneous skull and remaining inhomogeneities, and (5) model with homogeneous brain matter and remaining inhomogeneities. Scalp potentials because of three different dipolar sources in the parietal-occipital lobe were computed for all five models. Results of a forward solution revealed that tissues included in the model and the dipole source location directly affect the simulated scalp surface potentials. The major finding indicates that significant change in the scalp surface potentials is observed when the brain's distinctions are removed. The other modifications, for example, layers of the scalp and skull are important too, but they have less effect on the overall results.

In this paper, a system for speech-driven animation of generic 3D head models is presented. The system is based on the inversion of a joint Audio-Visual Hidden Markov Model to estimate the visual information from speech data. Estimated visual speech features are used to animate a simple face model. The animation of a more complex head model is then obtained by automatically mapping the deformation of the simple model to it. The proposed algorithm allows the animation of 3D head models of arbitrary complexity through a simple setup procedure. The resulting animation is evaluated in terms of intelligibility of visual speech through subjective tests, showing a promising performance.